1 Ligand Rearrangement and Hemilability in Rhodium(I) And
Total Page:16
File Type:pdf, Size:1020Kb
Ligand Rearrangement and Hemilability in Rhodium(I) and Iridium(I) Complexes Bearing Terphenyl Phosphines Juan José Moreno, María Fernández-Espada, Eric Krüger, Joaquín López-Serrano, Jesús Campos,* Ernesto Carmona. Instituto de Investigaciones Químicas (IIQ), Departamento de Química Inorgánica and Centro de Innovación en Química Avanzada (ORFEO-CINQA). Universidad de Sevilla and Consejo Superior de Investigaciones Científicas (CSIC). Avenida Américo Vespucio 49, 41092 Sevilla (Spain). Abstract We describe the synthesis of a series of cationic rhodium(I) and iridium(I) compounds stabilized by sterically demanding phosphines that contain a terphenyl substituent, PMe2Ar’ (Ar’ = 2,6-diarylphenyl radical). Salt metathesis of metal precursors [MCl(COD)(PMe2Ar)] (M = Rh, Ir; COD = cyclooctadiene) with NaBArF (BArF = B(3,5-C6H3(CF3)2)4) results in a series of cationic complexes in which the loss of the chloride ligand is compensated by the appearance of weak π-interactions with one of the flanking aryl rings of the terphenyl substituent. The same experiments carried out with carbonyl compounds [MCl(CO)2(PMe2Ar)] led to the corresponding cationic carbonyl complexes, whose CO-induced rearrangement reactivity has been investigated, both experimentally and computationally. The differences in reactivity between rhodium and iridium complexes, and as a result of varying the sterics of terphenyl phosphines are discussed. Introduction Phosphines are among the most widely used ancillary ligands in coordination and organometallic chemistry, largely due to the possibility of finely modulating their steric and electronic properties in a predictive manner.1 The range of physicochemical properties of transition metal complexes that can be tuned by the judicious choice of a phosphine ligand is exceptional. In particular, phosphines that contain bulky substituents have been successfully applied to stabilize a variety of low-coordinate transition metal complexes2 with important implications in catalysis.3 A notorious example is found in the work of Buchwald and coworkers, who synthesized in 19984 the first example of a currently extensive family of biaryldialkyl phosphines. The use of biaryl phosphines in homogeneous catalysis has found great success, particularly in palladium-catalysed cross coupling5 and amination reactions.6 Catalytic applications that rely on the combination of biaryl phosphines with iridium7 and rhodium8 precursors have also been described, although to a lesser extent. In an prominent example, Hartwig and co-workers demonstrated that the combination of [Rh(COD)2]BF4 (COD = 1,5- cyclooctadiene) and a biaryldialkyl phosphine functions as an excellent catalytic mixture for the intramolecular hydroamination of inactivated alkenes under mild conditions,9 with the active species containing the phosphine ligand bound to rhodium in a κ1,η6 coordination mode (A in Scheme 1).10 The group of Goldberg has investigated the chemistry of related rhodium complexes where the flanking aryl ring of several biaryl phosphines coordinates to the metal 1 centre with variable hapticity.11 These studies were later extended to the corresponding iridium systems.12 While rhodium designs were active catalyst for arene hydrogenation,11 their iridium counterparts revealed their potential for transfer hydrogenation reactions.12 More recently, the usefulness of Rh/biaryl phosphine combinations for the selective and catalytic functionalization of C-H bonds has been disclosed.13 + NMe2 R R Rh R R P PMe2 Cy Cy Xyl a, PMe2Ar 2 R = Me A Dipp i b, PMe2Ar 2 R = Pr Scheme 1. (a) Key intermediate A in the Hartwig’s Rh-mediated hydroamination of alkenes; (b) Terphenyl phosphine ligands employed in this work. The ability of biaryl phosphines to function as hemilabile ligands by means of dynamic π- arene coordination seems to be crucial for its success in catalysis.6a,14 Although this type of coordination enhances the stability of the active species, catalyst decomposition remains a major limitation, as observed with Goldberg11 and Hartwig’s10,15 systems. With this in mind, it is surprising that terphenyldialkyl phosphines,16 in which the biaryl substituent is replaced by a sterically more crowded terphenyl (2,6-diarylphenyl) group, have not been examined in deeper detail. We recently embarked to explore this type of ligands,17 demonstrating their potential to stabilize otherwise elusive organometallic frameworks.18 This work extends these studies to cationic rhodium and iridium systems bearing 1,5-cyclooctadiene and ethylene ligands. Thus, Xyl2 we analyse the coordination modes of two terphenyl phosphine ligands, namely PMe2Ar Dipp2 and PMe2Ar (Scheme 1). In addition, we have synthesized rhodium and iridium cationic derivatives where the olefins have been substituted by CO ligands, and have observed a CO- induced ligand rearrangement in which phosphine hemilability plays a key role. Results and Discussion Synthesis of Rh(I) and Ir(I) olefin compounds Xyl2 Dipp2 Salt metathesis of chloride complexes 1 (1a: Ar’ = PMe2Ar ; 1b: Ar’ = PMe2Ar ), cleanly prepared by the 1:1 reaction of [RhCl(COD)] and the corresponding terphenyl phosphine,17a with NaBArF (BArF = B(3,5-C6H3(CF3)2)4) in dichloromethane, afforded cationic complexes 2 quantitatively (Scheme 2). Chloride abstraction in compounds 1 triggers η6-coordination of the π-system of one of the flanking aryl substituents of the phosphine, forcing the COD ligand to change its bidentate η2:η2-coordination mode to monodentate η2-binding. Whereas there are examples of a phosphine aryl substituent coordinating to Rh(I) and Ir(I) in the η6-mode while maintaining the M—P bond,10,11,12 only a few complexes of these metals that exhibit monodentate COD ligand coordination have been reported. To our knowledge, prior to this work, no complexes of this kind had been structurally characterized for rhodium.19 2 Scheme 2. Synthesis of cationic rhodium complexes 2 by chloride abstraction from 1. The formulation proposed for complexes 2 in Scheme 2 was ascertained by NMR spectroscopy studies in solution and, in the case 2a, also by X-ray crystallography. In the 31P{1H} NMR 1 spectrum, a doublet at ca. 48 ppm, with a relatively large JPRh coupling of 192 Hz was observed for both 2a and 2b. Complex 2a shows fluxional behaviour in solution, as inferred from the 1 presence of some broad signals in the H NMR spectrum recorded at 25 ºC (500 MHz, CD2Cl2). Cooling at 0 ºC allowed freezing the structure pictured in Scheme 2 on the NMR time scale, thereby permitting assigning all the 1H and 13C NMR resonances in their respective spectra (see Supporting Information). Two of the aromatic proton resonances of compounds 2 appear 3 3 somewhat shielded at around 6.7 (d, 2 H, JHH ca. 7.5 Hz) and 5.9 ppm (t, 1 H, JHH ca. 7.5 Hz), in accordance with the proposed η6-coordination of one of the flanking aryl substituent of the terphenyl group. This is in contrast with corresponding resonances being found above 7.2 ppm for the analogous H atoms in the free flanking ring. In the 13C{1H} NMR spectrum of 2a, the associated 13C nuclei appear at 112.1 and 97.1 ppm for the Rh-bonded xylyl, and at 128.6 and 127.8 ppm for the non-coordinated xylyl substituent, and the same pattern is found for 2b. A similar dichotomy was found between these NMR parameters of the olefinic =CH groups of the COD ligand, with the proton and carbon nuclei of the Rh-bonded –CH=CH– moiety in 2 1 resonating at 3.38 and 74.0 ppm (d, JCRh = 13 Hz) respectively, while those corresponding to the uncoordinated C=C bond resonated at 5.55 and around 130 ppm. Dipp2 At variance with 2a, compound 2b based on the bulkier PMe2Ar phosphine led to sharp signals in its corresponding 1H NMR spectrum, with no sign of fluxionality on the NMR time scale. A plausible explanation for the dynamic behaviour of 2a could be the accomplishment of a fast dynamic equilibrium with unobserved amounts of an isomer featuring η2-coordination of the xylyl ring and bidentate η2:η2-binding of the COD ligand at 25 ºC. In fact, this latter structure was the preferred isomer in a related complex based on [(2- Biphenyl)dicyclohexylphosphine] (CyJohnPhos).11 Taking in mind the importance of these π- arene interactions in catalysis, it is interesting to note that the prevalence of a particular π- coordination mode in these rhodium complexes containing biaryl or terphenyl phosphines may be modulated by the choice of the phosphorus ligand. The proposed dynamic exchange between η2- and η6-coordination modes of the arene substituent is further supported by investigations on the analogous iridium system. Chloride abstraction by NaBArF from two iridium(I) compounds of formula [IrCl(COD)(PR2Ar’)] (3a: Ar’ = Xyl2 Dipp2 17a PMe2Ar , and 3b: Ar’ = PMe2Ar ) proceeded readily to yield cationic species 4 and 5. As depicted in Scheme 3 the equilibria between these complexes result from hapticity changes 3 within both the flanking arene of the terphenyl substituent (κ1 vs. η6) and the COD ligand (η2:η2 vs. η2). The equilibrium for 4a seems to be more rapid than in the rhodium analogue 2a, as inferred from the sharp resonances detected in the corresponding NMR spectra, and was confirmed by the presence of several exchange peaks in 2D EXSY experiments. The isomeric Xyl2 distribution is phosphine dependent, with the smaller PMe2Ar leading to isomer 4a as the Dipp2 major component (4a:5a; 59:41 ratio), whereas the more sterically congested PMe2Ar preferring COD η2, monodentate coordination accompanied by η6-coordination of the flanking Dipp ring (4b:5b; 14:86 ratio). Scheme 3. Synthesis of complexes 4 and 5 by chloride abstraction from compounds 3. The nature of isomers 4 and 5 was postulated on the basis of NMR spectroscopy, although the formulation of 4a was further confirmed by X-ray studies. Compounds 4 present a 31P{1H} resonance at around 15 ppm, slightly deshielded (ca.